Roger Sessions began a useful process Sept. 28 at his blog on the subject of IT project failures. On Oct. 2, I commented on his post (see Taking the Trillion-Dollar Figures out of the Costs of IT Failure) and Roger in turn commented as follows:
Thanks for helping to bring this problem to the forefront I don't mind people saying that I under-calculated, but it hurts the cause when people think I over-calculated.
After a few e-mails back and forth, I figured out that Roger thought I agreed with his multi-trillion-dollar estimate. I explained that I thought his estimate was way too high, but in case my blog post was not clear to others, let me try again. I do not agree with Roger's $6 trillion estimate of the cost of IT project failures annually. And I do not think of the issue of IT project failure as a "cause."
But I have decided it is a "problem." That's because I do believe his analytical approach to the issue has a lot of merit. I say that partially because I came up with a similar, but much less elaborate, analysis for my loosely related and purely coincidental Sept. 29 post on my IT Investment Research blog relative to IT project failure and the use of open source.
If you want to get right to what I think of the extent of the problem, skip the section below, "Applying the Roger Sessions Formula," which is all about calculations.
Applying the Roger Sessions Formula
I primarily disagree with Roger's main conclusion (he doesn't word it this way, but this is the logical outcome of his number) that almost every dollar spent on IT annually worldwide, around $6 trillion, is wasted because of IT project failure. That just flies in the face of everything I think I know about IT market dynamics from 40 years of research. I'm a firm believer in market forces, and that you can't fool all the people all the time.
To summarize, Roger and I both used the aggregate total of worldwide gross domestic products (GDP) in our formulas and we both mentioned Standish Group research in our blog posts. But after that, we differ as follows:
I also have some problems with the source of Roger's failure rate, at-risk rate and direct/indirect-cost ratios, but those numbers don't move the needle up to $6 trillion as much as the other numbers cited above.
Although I disagree with his conclusion, his formula is well thought out. So I suggest using the first part of Roger's formula, which involves calculating the amount of worldwide IT spend that is at risk of being involved in a failure. But I suggest using the total "6.4 percent of GDP" World Technology and Services Alliance number that he cites as a base so as to include personnel and other overhead costs. He uses only 43 percent of that number, saying he can't verify personnel and IT costs. I am not familiar with the World Technology and Services Alliance numbers he cites, but I found a Gartner source that is very similar and specifically includes personnel and overhead.
Remember, all of these numbers are just strawmen for the aggregate total of all your individual spending and theoretically failed spending, so when trillion dollar numbers are within a few hundred million, close enough!
I use the second half of Roger's formula, which attempts to actually calculate the cost of failed IT projects by multiplying some factors against a smaller World Technology and Services Alliance number that Roger says represents worldwide "hardware, software and services spending." I found an IDC number in the same ballpark.
Roger uses a number 10 times larger, the aggregate world GDP instead. I suggest that the lower number be used in the second half of his formula because we analysts and consultants often recommend that you in IT not make an IT investment unless "it will pay for itself in a year" in direct and indirect savings. Presumably, that is therefore what we analysts and consultants think the potential savings are in all IT project investments annually. I have never heard any analyst or consultant suggest you can save the entire OPEX of your company by doing a whole host of IT projects, which is how I interpret Roger's use of the aggregate world GDP in the way that he does.
The 10X difference in this part of the formula naturally leads to a different -- but I think more reasonable -- numerical outcome.
The Extent of the Problem
Again, remember that all the above numbers are just strawmen for the aggregate total of all your individual IT spending.
Using my changes to Roger's values, I come up with a number of around $500 billion as the annual cost of failed IT projects (about 20 percent of worldwide IT spending, based on Roger's estimate of IT spending equaling 2.75 percent of worldwide GDP, which I multiply against the World Bank's estimate of the aggregate world GDP. That number seems much more reasonable than over $6 trillion. As I said on this blog on Oct. 2, you:
... don't expect to save money out of your ongoing IT-sysadmin outsourcing expense when you are implementing a new purchasing application. You don't expect to reduce your firm's electricity bill by adding a new e-commerce feature to your Web site. Furthermore, many IT projects such as payroll processing are simply a cost of doing business, or involve meeting a government regulation such as Basel II or SOX, and so forth. There is no ROI at all expected of either (unless you want to calculate the fines you'll get from the various states for not paying your employees on time or from the SEC for not filing shareholder data in a timely manner).
So although I don't think the issue reaches the level of "a cause," if IT project failures are costing your enterprises 20 percent of your IT budgets annually, that's not chicken feed, either.
(I also recommend the thought Roger has put into what he calls on his blog "what I think is responsible for these (IT project) failures "
No matter what they cost the economy as a whole or what we analysts and consultants think of it, an IT project of yours that fails is costing you. And it's the number your top management will be thinking about.